diff --git a/doc/all-the-options.md b/doc/all-the-options.md
index f86f68f..c941c82 100644
--- a/doc/all-the-options.md
+++ b/doc/all-the-options.md
@@ -55,7 +55,7 @@ Any function that creates a `Statement` or a `PreparedStatement` will accept the
 
 * `:concurrency` -- a keyword that specifies the concurrency level: `:read-only`, `:updatable`,
 * `:cursors` -- a keyword that specifies whether cursors should be closed or held over a commit: `:close`, `:hold`,
-* `:fetch-size` -- an integer that guides the JDBC driver in terms of how many rows to fetch at once; sometimes you need to set `:fetch-size` to zero or a negative value in order to trigger streaming of result sets -- some JDBC drivers require additional options to be set on the connection _as well_,
+* `:fetch-size` -- an integer that guides the JDBC driver in terms of how many rows to fetch at once; the actual behavior of specifying `:fetch-size` is database-specific: some JDBC drivers use a zero or negative value here to trigger streaming of result sets -- other JDBC drivers require this to be positive for streaming and may require additional options to be set on the connection _as well_,
 * `:max-rows` -- an integer that tells the JDBC driver to limit result sets to this many rows,
 * `:result-type` -- a keyword that affects how the `ResultSet` can be traversed: `:forward-only`, `:scroll-insensitive`, `:scroll-sensitive`,
 * `:timeout` -- an integer that specifies the (query) timeout allowed for SQL operations, in seconds. See [**Handling Timeouts**](/doc/tips-and-tricks.md#handling-timeouts) in **Tips & Tricks** for more details on this and other possible timeout settings.